Digital Phenotyping, Remote Monitoring, and Risk Prediction of Heart Failure in Hypertensive Disorders of Pregnancy: A Systematic Review With Narrative Synthesis.
Abstract
Hypertensive disorders of pregnancy (HDPs), particularly preeclampsia, are associated with acute cardiovascular morbidity and future heart failure (HF). Current surveillance relies largely on blood pressure, symptoms, laboratory abnormalities, and clinician-triggered evaluation. Digital phenotyping, remote monitoring, electronic health record-based prediction, and machine learning may enable earlier cardiovascular risk identification. To systematically identify and narratively synthesize studies evaluating remote monitoring, digital health, machine learning, or risk-prediction approaches for HF-related phenotypes or cardiovascular morbidity in HDPs. MEDLINE, Embase, and Scopus were searched for studies published from January 2020 through April 10, 2025. Studies evaluating prediction or monitoring approaches in HDPs underwent sequential assessment for HF-related or cardiovascular morbidity outcomes. Screening and full-text reviews were performed by 2 independent reviewers, with disagreements resolved by a third reviewer. Given substantial heterogeneity, findings were synthesized narratively. Of 3366 records imported for screening, 6 studies met the inclusion criteria. Studies used clinical prediction models, electronic health record (EHR)-based models, and remote monitoring to evaluate acute cardiovascular morbidity, HFpEF, cardiac remodeling, and cardiovascular readmission. Model discrimination ranged from modest to strong (AUROC approximately 0.70-0.94). Recurrent predictors included severe hypertensive disease, earlier gestational age at delivery, kidney disease, pre-existing cardiac disease, pulmonary hypertension, maternal age, obesity-related measures, blood pressure, and postpartum volume retention. Most prediction models lacked external validation. Preliminary evidence suggests that clinical, EHR-derived, and remotely monitored data may support risk stratification for selected cardiovascular morbidity and HF-related phenotypes after HDPs. Prospective validation, physiological marker integration, and evaluation of risk-guided interventions are needed.
Abstract as supplied by the indexing source. © the authors / publisher.